Transverse Cirrus Bands (TCBs) are thin, streaky cloud patterns that often form high in the sky near the jet stream. These clouds are linked to dangerous air turbulence that can affect airplanes flying at cruising altitudes above 10 kilometers. Because TCBs are difficult to spot early with the human eye, they pose a serious risk to flight safety.

This study presents a new computer-based method that uses satellite images to automatically detect these clouds. The researchers used images from a Korean weather satellite called GEO-KOMPSAT-2A and trained a special kind of artificial intelligence model known as a Swin Transformer. This model is especially good at identifying the long and irregular shapes of TCBs.

To teach the model what TCBs look like, the researchers used real flight data from the International Air Transport Association and carefully labeled images where TCBs appeared. The model successfully detected TCBs with high accuracy and outperformed older detection methods.

This research shows how advanced AI can help weather monitoring systems better track turbulence-related clouds, which may improve flight safety and give pilots and airlines better warning systems in the future.